Ufdjrw

Tensorflow: 1.12

I am using bidirectional_dynamic_rnn. I wrote encoder_output to BahdanauAttention's memory option (recommended on tensorflow website), but it throws an error:

ValueError: Layer memory_layer expects 1 inputs, but it received 2
input tensors. Inputs received: tf.Tensor
'bidirectional_rnn/fw/fw/transpose_1:0' shape=(?, ?, 512)
dtype=float32, tf.Tensor 'ReverseSequence:0' shape=(?, ?, 512)
dtype=float32]

def model_inputs():
 inputs = tf.placeholder(tf.int32, [None, None], name='input')
 targets = tf.placeholder(tf.int32, [None, None], name='target')
 lr = tf.placeholder(tf.float32, name='learning_rate')
 keep_prob = tf.placeholder(tf.float32, name='keep_prob')
 return inputs, targets, lr, keep_prob

def preprocess_targets(targets, word2int, batch_size):
 left_side = tf.fill([batch_size, 1], word2int['<SOS>'])
 right_side = tf.strided_slice(targets, [0,0], [batch_size, -1], [1,1])
 preprocessed_targets = tf.concat([left_side, right_side], 1)
 return preprocessed_targets

#Encoder RNN
def encoder_rnn(rnn_inputs, rnn_size, num_layers, keep_prob, sequence_lenght):
 lstm = tf.contrib.rnn.BasicLSTMCell(rnn_size)
 lstm_dropout = tf.contrib.rnn.DropoutWrapper(lstm, input_keep_prob = keep_prob)
 encoder_cell = tf.contrib.rnn.MultiRNNCell([lstm_dropout] * num_layers)
 global encoder_output, encoder_state
 encoder_output, encoder_state = tf.nn.bidirectional_dynamic_rnn(cell_fw = encoder_cell,
 cell_bw = encoder_cell,
 sequence_length = sequence_length,
 inputs = rnn_inputs,
 dtype = tf.float32)

 return encoder_state



#Decoding training set
def decode_training_set(encoder_state, decoder_cell, decoder_embedded_input, sequence_lenght, decoding_scope, output_function, keep_prob, batch_size):
 #attention_states = tf.zeros([batch_size, 1, decoder_cell.output_size])
 attention_mechanism = tf.contrib.seq2seq.BahdanauAttention(num_units = decoder_cell.output_size, memory = encoder_output, normalize=False)

What can I do?